Object Scanning (iOS)

On iOS, a CameraObjectOutput scans codes, faces and bodies with the system’s built-in detectors, without any frame processing:

var codes by remember { mutableStateOf(emptyList<ScannedCode>()) }
val objectOutput = rememberObjectOutput(listOf(ScannedObjectType.QR, ScannedObjectType.EAN_13)) { objects ->
    codes = objects.filterIsInstance<ScannedCode>()
}
Camera(device = TargetCameraPosition.BACK, isActive = true, outputs = listOf(previewOutput, objectOutput))

Every ScannedObject has a type and a boundingBox:

  • A ScannedCode also has its decoded value and its cornerPoints.
  • A ScannedFace has its faceID and its rollAngle and yawAngle, when they are available.

Coordinates are in camera space (0.0 to 1.0). To draw them on the screen, convert the whole object with convertScannedObjectCoordinatesToViewCoordinates, on PreviewViewRef or CameraRef with Compose. It takes the preview’s scaling and cropping into account, and returns the same kind of object with its bounding box and corner points in view coordinates:

val viewCode = previewRef.convertScannedObjectCoordinatesToViewCoordinates(code) as ScannedCode

On Android, createObjectOutput throws, like in VisionCamera. Scan codes with the barcode scanner instead, which works on both platforms.


This site uses Just the Docs, a documentation theme for Jekyll.